university of birmingham for better or for worse?

University of Birmingham

For better or for worse?Hughes, Claire; Devine, R.T.

DOI:10.1111/cdev.12915

License:Other (please specify with Rights Statement)

Document VersionPeer reviewed version

Citation for published version (Harvard):Hughes, C & Devine, RT 2019, 'For better or for worse? Positive and negative parental influences on youngchildren's executive function', Child Development, vol. 90, no. 2, pp. 593-609. https://doi.org/10.1111/cdev.12915

Link to publication on Research at Birmingham portal

Publisher Rights Statement:Checked for eligibility: 20/09/2017

This is the peer reviewed version of the following article: Hughes, C. and Devine, R. T. (2017), For Better or for Worse? Positive andNegative Parental Influences on Young Children's Executive Function. Child Dev. doi:10.1111/cdev.12915, which has been published in finalform at 10.1111/cdev.12915. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving.

General rightsUnless a licence is specified above, all rights (including copyright and moral rights) in this document are retained by the authors and/or thecopyright holders. The express permission of the copyright holder must be obtained for any use of this material other than for purposespermitted by law.

•Users may freely distribute the URL that is used to identify this publication.•Users may download and/or print one copy of the publication from the University of Birmingham research portal for the purpose of privatestudy or non-commercial research.•User may use extracts from the document in line with the concept of ‘fair dealing’ under the Copyright, Designs and Patents Act 1988 (?)•Users may not further distribute the material nor use it for the purposes of commercial gain.

Where a licence is displayed above, please note the terms and conditions of the licence govern your use of this document.

When citing, please reference the published version.

Take down policyWhile the University of Birmingham exercises care and attention in making items available there are rare occasions when an item has beenuploaded in error or has been deemed to be commercially or otherwise sensitive.

If you believe that this is the case for this document, please contact [email protected] providing details and we will remove access tothe work immediately and investigate.

Download date: 24. Dec. 2021

https://doi.org/10.1111/cdev.12915

https://doi.org/10.1111/cdev.12915

https://birmingham.elsevierpure.com/en/publications/84a0a00d-f95e-4707-b771-b3a39e8ab5d4

Running Head: PARENTAL INFLUENCES ON CHILDREN’S EXECUTIVE FUNCTION

1

For Better or for Worse? Positive and Negative Parental Influences on Young

Children’s Executive Function.

Claire Hughes

Rory T. Devine

University of Cambridge

Author Note.

Centre for Family Research, Department of Psychology, University of Cambridge, Free

School Lane, Cambridge, CB23RQ, United Kingdom. We would like to thank Annabel

Amodia-Bidakowska, Giacomo Bignardi, Marie Constantinou, Hildigunnur Anna Hall,

Catherine Jones and Tessa Thwaites for their assistance with data collection and coding. This

study was funded by the UK Economic and Social Research Council (ES/JO21180/1) and the

Isaac Newton Trust, Cambridge. For correspondence concerning this article please e-mail

Claire Hughes: [email protected].

mailto:[email protected]


2



Abstract

Despite rapidly growing research on parental influences on children’s executive function

(EF), the uniqueness and specificity of parental predictors and links between adult EF and

parenting remain unexamined. This 13-month longitudinal study of 117 parent-child dyads

(60 boys; M Age at Time 1 = 3.94 years, SD = 0.53) included detailed observational coding

of parent-child interactions and assessed adult and child EF and child verbal ability (VA).

Supporting a differentiated view of parental influence, negative parent-child interactions and

parental scaffolding showed unique and specific associations with child EF, while the home

learning environment and parental language measures showed global associations with

children’s EF and VA.


3



Research on executive function (EF) is rooted in both cognitive and clinical psychology.

From a cognitive perspective, the term EF encompasses the higher-order processes that

underpin both goal-directed action and adaptive responses to novel, complex or ambiguous

situations (Hughes, Graham, & Grayson, 2005). Clinically, EF is used to describe the

functions of the prefrontal cortex (PFC), which differs from other brain regions in several

respects including protracted maturation extending well into adolescence (Kolb et al., 2012).

Measurements of EF typically centre on three related abilities: (1) ‘inhibitory control’ or the

ability to override entrenched habits or impulses; (2) ‘working memory’ or the ability to

update information held in mind; and (3) ‘set shifting’ or the ability to switch easily between

tasks (Garon, Bryson & Smith, 2008; Miyake & Friedman, 2012). In adolescence and

adulthood these abilities appear to be correlated but separable (Miyake & Friedman, 2012). In

early childhood however, individual differences in performance on a range of measures of EF

can be explained by a single latent factor (Wiebe et al., 2008). The explosion of interest in EF

in childhood over the past two decades reflects, at least in part, the recognition that individual

differences in EF underpin a range of important developmental outcomes. Preschool children

with advanced EF show better social understanding (e.g., Devine & Hughes, 2014), superior

academic ability (Devine, Bignardi & Hughes, 2016) and fewer behavioral problems (e.g.,

Hughes & Ensor, 2008). Understanding the origins of individual differences in EF is

therefore an important goal for developmental science.

Twin research highlights the potential importance of genetic influences on individual

differences in EF (e.g., Miyake & Friedman, 2012). Addressing this possibility, several recent

studies have investigated the intergenerational association in EF, with mixed results.

Specifically, in a study of 361 adopted 27-month-old children, Leve et al. (2013) found no


4

evidence for a correlation between children’s EF and individual differences their biological

parents’ performance on a conflict inhibition task. In contrast, Hughes & Ensor (2009)

reported a moderate correlation between parental planning and 4-year-old children’s EF that

persisted even when individual differences in verbal ability and age 2 EF were taken into

account. In two papers based on a study of 62 children across three time points (ages 2, 3 and

4), Cuevas and colleagues (Cuevas, Deater-Deckard, Kim-Spoon, Wang, Morasch, & Bell,

2014; Cuevas, Deater-Deckard, Kim-Spoon, Watson, Morasch, & Bell, 2014) also reported

a moderate and significant association between parental EF and individual differences in

children’s EF at ages 2, 3 and 4. These cross-sectional associations held even when individual

differences in children’s family background, age and verbal ability were taken into account.

However, when effects of prior levels of child EF and observed parental sensitivity were

controlled, parental EF did not make a unique contribution to children’s EF at age 4.

Although it is difficult to tease apart effects of nature and nurture without a

genetically sensitive design (e.g., adoption study), there are guiding principles that emerge

from research on behavioral genetics. For example, two of the ‘big ideas’ offered by Asbury

and Plomin (2014) in their review of the literature on genetic influences on education, are

that: (1) genetic factors underpin temporal continuity while environmental factors contribute

to change over time; (2) genes are generalists (e.g., the same genes contribute to individual

differences in both anxiety and depression – Eley & Stevenson, 1999) and environments are

specialists (i.e., people with the same genetic susceptibility to anxiety or depression may

manifest very different symptoms, depending on the nature of environmental adversity

experienced). Each of these conclusions leads to a testable hypothesis. For example, in line

with the ‘genes are generalists’ view, one would expect parent EF to be correlated with a

range of child cognitive abilities (e.g., both EF and verbal ability); in turn, this leads to the

hypothesis that the intergenerational association in EF would be attenuated once other more


5

general aspects of children’s cognitive ability are controlled. Second, if genetic factors

underpin stability in any given trait, controlling for children’s prior EF ability should

attenuate any genetically mediated association between parent and child EF.

In contrast with the genetic perspective outlined above, a growing body of research

indicates that the development of EF, like language, but unlike many other neuro-cognitive

functions, is particularly susceptible to environmental influence (Noble, Norman, & Farah,

2005). For example, EF in childhood is robustly associated with variation in socio-economic

status (Ursache & Noble, 2016) and impaired in post-institutionalized children (Merz, Harlé,

Noble, & McCall, 2016). Likewise, longitudinal research using latent growth models has

demonstrated that both early exposure to maternal symptoms of depression and persistence of

depressive symptoms show unique associations with delays in children’s EF between the ages

of 2 and 6 (Hughes, Roman, Hart, & Ensor, 2013). In short, research evidence suggests that

families can both help and hinder preschool children’s developing EF skills: socio-cultural

accounts of cognitive development emphasize the benefits of support from more expert social

partners (e.g., Vygotsky, 1978), while clinical research evidence highlights the adverse

effects of stressful environments on neurocognitive development (e.g., Blair & Raver, 2015).

Historically, these two perspectives have been adopted by different research groups

such that that these distinct accounts have yet to be integrated to examine positive and

negative parental predictors of preschool children’s EF in tandem. This is a notable omission

given the growing evidence for specific links between different parental characteristics and

specific child outcomes (e.g., Smetana, 2017). In this ‘domains’ approach, socialization is

viewed not as a general process but as a collection of distinct processes that act on specific

domains of development (Davidov, 2013). Adopting this differentiated view of parental

influence (Grusec & Davidov, 2010), the current study aimed to integrate clinical and socio-

cultural accounts by examining the uniqueness and overlap of both positive and negative


6

parental influences on individual differences in EF in a longitudinal study of 117 preschool

children and their parents.

Do Both Positive and Negative Dimensions of Parental Behavior Influence Young

Children’s EF?

While evidence from large-scale longitudinal studies demonstrates that parental

sensitivity predicts developmental gains in EF (Blair, Raver, & Berry, 2014), findings from

two more detailed studies highlight the likelihood that multiple independent processes

contribute to the association between family environment quality and young children’s EF.

Specifically, Hughes and Ensor (2009) found that gains in EF between the ages of 2 and 4

showed independent associations (in the expected directions) with family chaos, the richness

of parental language, verbal scaffolding (as measured using a global rating of parental praise,

elaboration and encouragement) and parental EF. Likewise, Bernier, Carlson and Whipple

(2010) showed that measures of maternal sensitivity, mind-mindedness and autonomy

support (a global rating of parental support and flexibility during structured play) gathered at

12 to 15 months each correlated with child EF at 26 months. Moreover autonomy support

predicted EF at 26 months even when individual differences in early cognitive ability were

taken into account. Importantly, while the two studies above were unusual in including

measures of parental scaffolding, sensitivity and support, neither incorporated explicit

measures of children’s responses to parents’ interventions. This is a notable omission for at

least two reasons. First, in a review of the efficacy of a wide variety of school-based

interventions designed to promote EF development, Diamond and Lee (2011) noted that a

carefully monitored progression in level of difficulty or challenge was one of the key

ingredients for success. Second, Wood and colleagues’ original and seminal research on

mothers’ tutoring strategies (Wood, Bruner, & Ross, 1976; Wood & Middleton, 1975)


7

showed that effective ‘scaffolding’ hinges upon the parent’s ability to modify their level of

support in a way that is contingent upon children’s difficulty or success on the task.

This process of ‘contingent shifting’ (Wood et al., 1976) involves the parent (or expert

partner) following a contingency rule, altering the specificity of instruction to provide direct

instruction when needed while gradually transferring responsibility as the child shows

increasing task mastery. In other words, research on scaffolding suggests that alongside

attention, warmth, stimulation and encouragement, parents also need to demonstrate

flexibility in order to promote children’s goal-directed activities. The ability to be flexible in

feedback in order to follow the contingency rule has, as one might expect, been found to be

higher in parents rated as authoritative (e.g., Pratt, Green, MacVicar, & Bountrogianni, 1992)

or sensitive (Mulvaney, McCartney, Bub, & Marshall, 2006). Underscoring the importance of

parent-child contingency, studies incorporating measures of parent-child contingency during

problem-solving tasks demonstrate that individual differences in parental scaffolding show

longitudinal relations with children’s EF (Hammond, Muller, Carpendale, Bibok, &

Lieberman-Finestone, 2012).

Broadly speaking, existing studies have shown that individual differences in

children’s EF are related to both the quality and quantity of affective and cognitive support

that parents provide. However, as noted in a recent review of research on parental influences

on young children’s EF (Fay-Stammbach, Hawes & Meredith, 2014), it is unclear whether

these different aspects of parental support show independent or overlapping associations with

child EF. Thus in the current study we included several different direct observational ratings

of parental behavior. Integrating the socio-cultural and clinical traditions, we included

quantitative measures of family learning support alongside direct measures of the cognitive

and affective quality of parent-child interactions to investigate the unique relations between

these distinct aspects of parental behavior and children’s EF. We predicted that parental


8

scaffolding (as measured by contingent shifting) and negative parent-child interactions (i.e.,

interactions characterized by negative affect and parental control) would exhibit unique and

opposing associations with variation in children’s EF.

Do Parental Influences on Children’s EF Overlap with Parental Influences on other

Cognitive Abilities?

With very few exceptions (e.g., Hughes & Ensor, 2009), previous studies have not

reported on the specificity of parental influences on EF. Do parental behaviors, such as

scaffolding, contribute to variation in EF in particular or to children’s cognitive skills more

generally? Answering this question is crucial to adopting a ‘domains’ approach to the

socialization of EF (e.g., Davidov, 2013). We therefore examined the relations between each

of the parental behaviors in our study and growth in children’s verbal ability. Verbal ability,

like EF, is known to be related to a range of parental factors, such as socio-economic status

and early parent-child interactions (Hoff, 2006). One possibility, discussed in Devine,

Bignardi and Hughes (2016), is that there are domain-general effects of parental cognitive

support on both EF and verbal ability. For example, developmental associations have been

reported between parental scaffolding and children’s later academic success and adjustment

(e.g., Pratt, Green, MacVicar, & Bountrogianni, 1992). Equally, the provision of informal

learning opportunities (or the ‘Home Learning Environment’), a well-recognized predictor of

children’s language skills (e.g., Son & Morrison, 2010), might also be expected to provide a

training ground for children’s developing EF. For example, as noted by Blair and Raver

(2015), reading activities require children to shift their attention between phonemes and

whole words. Likewise, the richness of children’s linguistic environments is known to predict

variation in children’s language skills (e.g., Hoff, 2006) and so may indirectly foster the

development of children’s EF. On the other hand, there might also be domain general effects

of parental affective support. For example, parent-child interactions characterized by conflict


9

and control have been shown to impact on children’s language development (MacKenzie,

Nicklas, Waldfogel & Brooks-Gunn, 2012). The domain-general account leads to the

prediction that each aspect of parental behavior will be of similar importance for both

children’s EF and VA. In contrast, a domain-specific account predicts that there will be

particular aspects of parenting that are differentially salient for either child EF or child VA.

More specifically, while both parental scaffolding (as measured by contingent shifting) and

negative parent-child interactions (i.e., interactions characterized by negative affect and

parental control) are hypothesized to exhibit unique associations with variation in children’s

EF, parental language input is hypothesized to be specifically related to individual differences

in children’s verbal ability. Identifying what aspects of parental behavior shape individual

differences in EF (and not both EF and VA) will shed light on the potential developmental

mechanisms involved in the socialization of EF.

Summary of Study Aims

Overall then, the evidence from existing research suggests that individual differences

in preschoolers’ EF are associated with variations in: parental EF, parental cognitive support

(e.g., scaffolding or ‘contingent shifting’); the affective quality of parent-child interactions

(e.g., negative parent-child interactions); and more general features of the family environment

(e.g., parental language input and home learning opportunities). Unlike previous studies, the

current investigation encompasses each of these potential parental influences upon EF

development in the preschool years. In addition, by conducting parallel analyses with EF and

VA as outcome variables, we aimed to assess the specificity of parental influences on each

cognitive domain. In sum, this study had two primary goals: (1) to unite previously disparate

bodies of research by studying both positive and negative parental influences on children’s

EF in order to establish whether individual parental measures show unique or overlapping


10

associations with child EF; and (2) to conduct parallel analyses with child VA as the

outcome, in order to assess the specificity of parental influences on child EF.

Method

Participants

Participants were recruited from nurseries, shopping centres and playgroups in the

East of England. Children had to be aged either 3 or 4 years old without a history of

developmental delay and be native English speakers. One hundred twenty parent-child dyads

took part at Time 1as part of a larger international study of children’s socio-cognitive

development (Hughes, Devine & Wang, 2017). One hundred seventeen dyads (60 boys)

agreed to be contacted for a follow-up study (3 families planned to leave the region). Families

were socio-economically homogenous (81% of parents had a degree level education) but

ethnically diverse (66% White British). Of the 117 families contacted at Time 2, two were no

longer eligible to participate as they had left the country. Of the eligible 115 families, 103

(90%) took part at Time 2 approximately 13 months later, SD = 1.65 months, range: 11 - 17

months. At Time 1 children were aged 3.94 years, SD = 0.53, range: 3 – 4.95 years, and at

Time 2 children were aged 5.11 years, SD = 0.54, range: 4 – 6.10 years. Non-returners did

not differ from those who returned for the second visit in age, gender or general cognitive

ability, but were marginally more likely to have low levels of parental education, OR = 3.05,

B = 1.12, SE = 0.64, p = .08.

Procedures

The University Research Ethics Committee approved all procedures prior to

commencement of the study. Parents and children participated in two laboratory visits lasting

up to 75 minutes approximately 1 year apart. Children completed tasks designed to measure

EF and general cognitive ability. Individual child cognitive testing lasted approximately 30

minutes. Child completed the task battery in a fixed counter-balanced format such that no two


11

tasks from any domain (e.g., verbal ability, EF) were completed in succession. Children were

allowed rest breaks and received stickers for completing each task. Parents completed a

questionnaire booklet and computerized EF tasks in another room while children underwent

cognitive testing. Parents were then observed interacting with their child during 5 minutes of

structured play with a set of jigsaw puzzles and 5 minutes of free play. Parents were

debriefed and provided with £15 and a small gift for their child. Children completed three

measures of EF at both Time 1 and 2. Two additional tasks were administered at Time 2. The

DCCS Border Game was not included at Time 1 because pilot work indicated that the

youngest children would find the task too difficult and the Day/Night Stroop was not

included at Time 1 to minimize problems of fatigue. Adding these two tasks to the EF

battery when the children were older and able to concentrate for a longer period had the

additional advantage of providing sufficient time for a second researcher to administer a

parallel parent EF battery at Time 2.

Measures

Child EF. Children completed a short battery of tasks designed to measure individual

differences in EF at Time 1 (March – September, 2013) and 2 (March – September 2014).

These represent widely-used and valid measures of EF in early childhood (Garon, Bryson &

Smith, 2008). The children completed the Happy-Sad Stroop (Lagattuta, Sayfan, & Monsour,

2011) at both time points as a measure of conflict inhibition. Children were shown two 8cm x

8cm cards depicting either a yellow ‘happy face’ or a yellow ‘sad face’. Children were asked

to point to the happy face and then to the sad face. The examiner then told the children that

they would play a ‘silly game’ so that when the examiner said ‘happy’ the child had to point

to the sad face and when the examiner said ‘sad’ the child had to point to the happy face.

Children completed 4 training trials with feedback from the examiner and received up to two

further sets of 4 training trials if any errors were made. If unsuccessful on training trials,


12

children were assigned a score of 0 and testing was discontinued. In the test phase of the task

the children completed 20 trials in a fixed pseudo-random order with no feedback. The

number of correct items was summed together.

Children completed the Dimension Change Card Sort (DCCS) Task (Zelazo, 2006) at

both Time 1 and 2 as a measure of their ability to switch flexibly between rules. In each

phase the children were shown two 10cm x 7.5cm target cards (one showing a blue rabbit and

the other a red boat) attached to two sorting boxes and were required to sort six cards

depicting either a blue boat or red rabbit. The pre-switch phase consisted of either six trials of

the ‘color game’ (i.e., matching the sorting card to the color of the target card) or six trials of

the ‘shape game’ (i.e., matching the sorting card to the shape depicted on the target card)

(counter-balanced across participants). All children passed the pre-switch phase (i.e., sorted 5

or more cards correctly) and those children playing the color game proceeded to the shape

game and vice versa. In the post-switch phase the children were told that the rules had

changed and proceeded to either the shape game or color game depending on what game they

had completed in the pre-switch phase. Children were awarded 1 point for each correctly

sorted card in the post-switch phase. At Time 2 children passing the post-switch phase (i.e.,

sorted 5 or more cards correctly) completed the border game. In the border game children

sorted a further 12 cards containing 6 normal sorting cards and 6 cards with a thick black

border. Children sorted the normal cards according to one rule (e.g., shape game) and the

black border cards according to another (e.g., color game) receiving 1 point for each correctly

sorted card. Children who failed the post-switch phase received 0 on the border game.

Children completed the Self-Ordered Pointing Task (Cragg & Nation, 2007) at Time

1 and 2 as a measure of working memory. Children were shown a color flipbook with an

increasing number of pictures of single syllable objects (ranging from 2 objects to 7 objects

with two sets in each number) in one of 16 locations on the page. For example, the first page


13

depicted two images (e.g., bowl, flag) and the second page depicted the same images but in

different spatial locations. No two objects were taken from the same class of objects (e.g.,

fruits, toys, pets). Children were required to point to a new picture on each page and could

not select the same picture twice. The task began with two demonstration sets (with

experimenter feedback). Repetition errors (i.e., repeated points to the same picture) were

recorded and used as an index of working memory. These scores were reflected for

consistency with other measures of EF in the study (i.e., high scores indicated greater EF).

Children also completed the Day/Night Stroop (Gerstadt, Hong, & Diamond, 1994) at

Time 2 to measure of conflict inhibition. This task followed the same procedure as the

Happy-Sad Stroop using two 10cm x 7.5cm cards depicting either the sun or the moon

instead of cards depicting happy and sad faces. Children pointed to the picture of the sun

when the examiner said ‘night’ and to the picture of the moon when the examiner said ‘day’.

Children completed 20 trials in a fixed order receiving 1 point for each correct trial.

Child Cognitive and Language Ability. Children completed two tasks from the

Wechsler Preschool and Primary Scale of Intelligence (WPPSI-III-UK) (Rust, 2003).

Children completed the Object Assembly task at Time 1to measure General Cognitive Ability

(IQ). After two training trials in which the examiner demonstrated how to use the puzzle

pieces, children assembled a set of two-dimensional puzzles depicting cartoon images of

objects (e.g., clock, bird). The number of correctly aligned junctures in the first 90s of trial

was recorded. The children completed up to 14 trials and scores from each trial were summed

together. Testing was discontinued if children scored 0 on three consecutive trials. The

Object Assembly task is strongly correlated with Full Scale and Performance Intelligence

Quotient in young children and shows excellent test-retest reliability (Wechsler, 2002).

Children completed the Receptive Vocabulary task at Time 1 and 2 to measure Verbal

Ability. Children were asked to point to a picture shown in a color flipbook depicting 4


14

images on each page that matched a word uttered by the examiner. Children received 1 point

for each correctly identified picture and testing was discontinued if children made 5

consecutive errors. Children completed up to 38 trials and scores from each item were

summed together. The Receptive Vocabulary subscale of the WPPSI shows excellent test-

retest reliability and is strongly correlated with Verbal Intelligence Quotient in young

children and so provides a valid index of children’s verbal ability (Wechsler, 2002).

Negative Parent-Child Interaction, Scaffolding and the Home Learning

Environment. Parents and children were recorded at Time 1 for 5 minutes playing together

using wall-mounted digital cameras. Parents and children were provided with three jigsaw

puzzles (a 6-, 8- and 12-piece puzzle) from the Galt Velvet Puzzles Jigsaw set. Parents were

asked to work with their child to complete as many of the three puzzles as possible within 5

minutes. The task was based on procedures used in previous studies of individual differences

in parent-child interactions and parental scaffolding (e.g., Carr & Pike, 2012). Observed

parental behaviors in similar tasks show stability over time and across contexts (e.g., Hughes

& Ensor, 2007). The data from these videos were then coded off-line using two different

coding schemes by different trained researchers naive to the participants’ test scores.

Negative Parent-Child Interaction was assessed using the Parent-Child Interaction

System (PARCHISY) coding scheme (Deater-Deckard, Pylas, & Petrill, 1997). Parents

received scores on three 7-point rating scales (ranging from ‘none’ to ‘exclusive/constant’):

negative control (i.e., use of physical control, use of criticism), negative affect (i.e., frowning,

harsh tone of voice) and conflict (i.e., disagreement, arguing or tussling). Double coding of

25 randomly selected videos showed acceptable intra-class correlations for each item:

negative content ICC = .89, negative affect ICC = .75, and conflict ICC = .74. The remaining

clips were double-coded and scores were averaged across raters.


15

We measured Parental Scaffolding using a coding scheme devised by Wood and

Middleton (1975) and later revised by Meins (1997). Each verbal and non-verbal task-related

behavior was rated by coders naïve to parents’ scores on Negative Parent-Child Interaction.

We assigned parental behaviors into one of five mutually exclusive categories: Level 1

Orienting Verbal Suggestions (e.g., ‘Let’s start with the corners’); Level 2 Suggestions about

Specific Pieces or Locations or Actions (e.g., ‘Try turn that piece around’); Level 3 Verbal

Solutions (e.g., ‘This piece goes here’); Level 4 Direct Physical Solutions (e.g., Caregiver

hands child a piece for a specific location); Level 5 Physical Demonstrations (e.g., Caregiver

assembles or dismantles parts of the puzzle) (Devine, Bignardi & Hughes, 2016). We coded

children’s responses as either a ‘success’ (i.e., correct placement of the puzzle piece) or a

‘failure’ (i.e., incorrect placement of the piece). ICCs derived from 25 randomly selected

double-coded videos, were acceptable for all codes: Level 1 intervention frequency ICC =

.64, Level 2 intervention frequency ICC = .85, Level 3 intervention frequency ICC = .97,

Level 4 intervention frequency ICC = .98, Level 5 intervention frequency ICC = .96,

frequency of child successes ICC = .99 and frequency of child failures ICC = .94. The

relatively low level of inter-rater agreement for the Level 1 interventions was accounted for

by two cases where the two raters differed by more than 2. Removal of these cases increased

the ICC to .87.

We organised parent-child codes into three-turn chains of parent interventions, child

actions and parent responses. Where numerous interventions preceded a child action we

selected the highest level of intervention (Carr & Pike, 2012;Wood & Middleton, 1975). We

used these three-turn chains to assess the contingency between parents and children during

the activity. We recorded the amount of times that parents shifted ‘up’ (i.e., moving from a

less specific to more directive intervention level), shifted ‘down’ (i.e., moving from directive

to less specific intervention level) and remained at the same level of intervention (‘no shift’)


16

after each child response. Parental scaffolding was measured through parents’ use of the

contingency rule (Carr & Pike, 2012; Meins, 1997; Wood & Middleton, 1975), whereby a

child’s success should be followed by an intervention at the same or at a lower level of

specificity and a child’s failure should be followed by an intervention that is one or two levels

higher than the previous level of intervention. We calculated contingency or scaffolding

scores by summing the total number of times that parents used the contingency rule and

dividing this by the total number of parental interventions after each success or failure.

At Time 2 parents completed Home Learning Environment (HLE) Index (Melhuish et

al., 2008). Parents reported on the frequency with which they engaged in seven activities with

their children (e.g., reading at home, teaching numbers and counting) on a 7-point scale

(ranging from ‘occasionally or less than once a week’ to ‘7 times a week/constantly’). Parents

received a score of 0 if they did not engage in a particular activity with their child. Internal

consistency of this measure was acceptable (α = 0.73) and so the scores from each item were

summed together.

Parent Language Input and EF. Parental Mean Length of Utterance (MLU) was

measured using verbatim transcripts of parent-child talk during 5-minutes of free play at

Time 1 to provide an index of parental language input. Parents and children were recorded for

5 minutes playing together using wall-mounted unobtrusive digital cameras while the

experimenters were in another room. Parents and children were provided with the Hasbro

Play-Doh Sweets Lunchbox play set and instructed to play together as they normally would.

Measures of parent-child talk during such observations are highly stable across time and

across different observational contexts (Ensor, Devine, Marks & Hughes, 2014). The

recordings were transcribed verbatim and analysed using Linguistic Enquiry and Word Count

(LIWC) software (Tausczik & Pennebaker, 2010) to ascertain the parents’ mean length of

utterance by dividing the total number of words spoken by the total number of parental


17

utterances during the 5-minute observation. We manually coded 25% of the transcripts from

these interactions and found high levels of inter-rater reliability between a human coder and

the LIWC software, ICC = .93.

Parental EF was measured using two computerized tasks at Time 2 based on tests

designed by Davidson, Amso, Anderson, and Diamond (2006) to measure simple inhibition

and task switching. In the Arrows Task (designed to measure inhibition) parents viewed one

of four images of a purple arrow on a white screen. The arrow pointed either directly

downward or diagonally on either the left- or right-hand side of the screen. During the

congruent (control) trials, the arrow pointed directly downward and participants had to press

the button on the same side as where the arrow appeared. During the incongruent (test) trails,

the arrow pointed diagonally and participants had to press the button on the opposite side to

where the arrow appeared. Before beginning participants received detailed instructions and

practice items. Each 750-ms trial was preceded by a 250-ms interval in which a 16-point font

black crosshair was displayed against a white background. Participants had to respond with a

button press within the trial period to each of 20 control trials and 20 test trials administered

in a random order. Efficiency scores were based on the total number of incongruent trials

divided by the total time taken on incongruent trials. In line with previous studies anticipatory

responses (<200ms) and missed responses were recorded as errors (Davidson et al., 2006).

The Hearts and Flowers test (designed to measure task switching) consisted of three

separate tasks. Each task consisted of 20 trials in which one of two pictures (a red heart or red

flower) was presented for 750-ms on either the left- or right-hand side of a white screen. Each

trial was preceded by a 250-ms interval in which a black cross-hair was presented centrally

on a white screen. In the congruent condition parents had to respond as quickly as possible by

hitting a button on the same side as the red heart. In the incongruent condition parents had to

respond as quickly as possible by hitting a button on the opposite side to where the red flower


18

appeared. In the final mixed condition parents were had to press a button on the same side

when a heart appeared and on the opposite side when a flower appeared. Parents completed

each condition in a fixed order but trials within each condition were presented in a random

order. Efficiency scores were calculated for the mixed condition by dividing the total number

of correct responses by the total time taken on mixed trials. In line with previous studies

anticipatory responses (<200ms) and missed responses were recorded as errors (Davidson et

al., 2006).

Results

Analytic Strategy

We conducted our analyses in MPlus Version 7 (Muthèn & Muthèn, 2012) using a

robust maximum likelihood estimator which is appropriate when data are non-normally

distributed data and sample sizes are small (Brown, 2015). We used a full information

approach to data analysis so that all cases with data at Time 1 were included in the analyses.

Missing values were estimated in MPlus using the robust maximum likelihood estimator.

This provides reliable estimates of missing values based on participants’ scores on other

variables in the model (Asparouhov & Muthen, 2010). For our primary analyses we used an

auto-regressive structural equation model to examine the relations between parental measures

and children’s EF and verbal ability (VA) at Time 2. In auto-regressive models the dependent

variable (X2) is regressed onto an earlier measure of that same variable (X1) and the predictor

variables of interest (Y1) (Hertzog & Nesselroade, 2003; Menard, 2002). The predictor

variable (Y1) is said to have a lagged or distal effect on the dependent variable (X2) if it

shows a significant association with the dependent variable when earlier scores on the

dependent variable (X1) are taken into account (Menard, 2002). We used Brown’s (2015)

four criteria to evaluate model fit: a non-significant χ2 test of model fit, comparative fit index


19

(CFI) ≥ 0.95, Tucker Lewis index (TLI) ≥ 0.95, and root mean square error of approximation

(RMSEA) ≤ 0.08.

Data Reduction and Descriptive Statistics for Children’s Measures

Descriptive statistics for all parent and child measures are shown in Table 1.

Children’s performance improved significantly on each of the repeated measures of EF.

Specifically children were more accurate on the Happy-Sad Stroop, MDiff = 3.93, SDDiff =

5.78, 95%CI [2.75, 5.10], t (102) = 6.63, p < .001, sorted more cards correctly in the post-

switch phase of the DCCS, MDiff= 1.62, SDDiff = 2.58, 95%CI [1.12, 2.13], t (102) = 6.37, p <

.001, and made fewer errors as a proportion of total trials on the SOPT, MDiff = 0.05, SDDiff =

0.06, 95%CI [0.03, 0.06], t (102) = 7.28, p < .001. Correlations between measures of EF were

moderate at Time 1, .29 < r < .48, Mean r = .40, and weak to moderate at Time 2, .08 < r <

.73, Mean r = .33. The SOPT error score at Time 2 was not correlated with any other measure

of EF at Time 2 and so was not included in any further analyses. As reported elsewhere

(Devine, Bignardi & Hughes, 2016), we tested a measurement model in which each EF

indicator at Time 1 loaded onto one latent factor and each EF indicator at Time 2 loaded onto

a separate latent factor. The error terms for the two DCCS indicators at Time 2 were

permitted to correlate. This model provided an acceptable fit to the data, χ (12) – 16.47, p =

0.17, CFI = 0.97, TLI = 0.96, RMSEA = 0.06. All indicators loaded significantly (all ps <

.01) onto the Time 1 EF latent factor, Standardized Estimates: .50 – .78, and all but one

indicator (i.e., the DCCS Post-Shift score) loaded significantly onto the Time 2 EF latent

factor, Standardized Estimates: .24 - .60. The factor determinacy co-efficient was 0.87 for the

Time 1 latent factor and 0.84 for the Time 2 latent factor indicating high internal consistency

(Brown, 2015). EF scores were therefore standardized and averaged to create a single EF

variable for each time point.


20

There were strong concurrent and longitudinal correlations between each of the child

measures (Table 2). Individual differences in EF and VA showed strong rank-order stability

across time. EF at Time 1 remained significantly correlated with EF at Time 2 even when

general cognitive ability and age (at Time 2) were controlled, pr (99) = .26, p < .01. VA at

Time 1 remained correlated with VA at Time 2 even when age (at Time 2) and general

cognitive ability were taken into account, pr (99) = .45, p < .001.

Data Reduction and Descriptive Statistics for Parents’ Measures

Parents’ performance on the measures of EF conformed to expectations. That is,

parents were more efficient in their performance on the congruent trials, M = 1.47, SD = 0.61,

95%CI [1.35, 1.59], of the Arrows Task than on the incongruent trials, M = 1.14, SD = 0.57,

95%CI [1.02, 1.25], t (102) = 9.68, p < .001. Likewise there was a main effect of condition

on performance on the Hearts and Flowers task, F (2, 204) = 592.38, p < .001. Efficiency

scores were highest in the congruent condition, M = 2.64, SD = 0.46, 95%CI [2.55, 2.74],

followed by the in the incongruent condition, M = 2.03, SD = 0.56, 95%CI [1.92, 2.14], and

worst in the mixed condition, M = 0.99, SD = 0.54, 95%CI [0.89, 1.11], all contrasts were

significant, ps < .01. Efficiency scores from both the Arrows Task (incongruent condition)

and the Hearts and Flowers Task (mixed condition) were strongly correlated, r (101) = .60, p

< .001, and so scores from each task were standardized and averaged together to create a

single Parent EF variable (α = .75).

Each of the PARCHISY items (i.e., Negative Control, Negative Affect and Conflict)

were significantly inter-correlated, .36 < r < .54, all ps < .01. As reported elsewhere (Devine,

Bignardi & Hughes, 2017) we tested a single latent factor measurement model in which each

of these items loaded onto a ‘Negative Parent-Child Interaction’ factor. This latent factor

exhibited significant variance, unstandardized estimate = 0.45, p = .007, and all item loadings

were significant: Negative Control standardized estimate = .52, p < .001, Negative Affect


21

standardized estimate = .68, p < .001, Conflict standardized estimate = .80, p < .001. The

factor determinacy co-efficient was 0.87 indicating high internal consistency (Brown, 2015).

PARCHISY items were therefore standardized and averaged to create a single ‘Negative

Parent-Child Interaction’ score.

Main Analysis

Relations between parental behavior and children’s EF and VA. We used

structural equation modelling to investigate the uniqueness and specificity of the relations

between each of the parental measures and individual differences in children’s EF and VA at

Time 2. We specified a longitudinal structural equation model with two correlated dependent

variables: EF at Time 2 and VA at Time 2. Using an auto-regressive approach we regressed

each dependent variable onto earlier scores on that variable at Time 1. We also regressed

each of the dependent variables onto general cognitive ability at Time 1, SES and age at Time

2 to control for the effects of these variables. Finally we permitted each of the predictor

variables to co-vary in the model.

This model provided an acceptable fit to the data, χ2 (1) = 0.56, p = .45, CFI = 1.00,

TLI = 1.07, RMSEA = 0. The model accounted for 48% of the variance in EF at Time 2 and

53% of the variance in VA at Time 2. To aid ease of interpretation only significant

standardized parameter estimates for this longitudinal model are presented in Figure 1. The

complete results for this model are presented in Table S1. Several findings from this model

deserve note. There were general effects of the HLE on both EF at Time 2, β = .14, p < .05,

95%CI [.03, .25], and VA at Time 2, β = .17, p < .05, 95%CI [.04, .30]. Likewise there were

general effects of parental MLU on later VA at Time 2, β = .21, p < .01, 95%CI [.07, .35],

and on EF at Time 2, β = -.14, p < .05, 95%CI [-.25, -.03], but in opposite directions. Parent

EF did not independently predict either child EF at Time 2, β = .09, p =.19, 95%CI [-.03,

.22], or VA at Time 2, β = .004, p = .95, 95%CI [-.10, .10], once all other measures were


22

taken into account. Negative parent-child interactions, β = -.23, p < .01, 95%CI [-.36, -.10],

and parental scaffolding, β = .19, p < .01, 95%CI [.09, .29], showed unique specific

associations with EF at Time 2 (each accounting uniquely for 4% of the variance in EF at

Time 2) but not with VA at Time 2.

Uniqueness and specificity of the relations between parental behavior and

children’s EF and VA. To investigate whether the associations between parental negative

parent-child interactions, parental scaffolding and EF at Time 2 were significantly different

from the relations between these parental variables and VA at Time 2, we specified a model

in which we constrained the regression paths between negative parent-child interaction and

EF at Time 2 and negative parent-child interaction and VA at Time 2 to be equal. We also

constrained the regression paths between parental scaffolding and EF at Time 2 and parental

scaffolding and VA at Time 2 to be equal. These constraints resulted in a significant decrease

in model fit, χ2 (3) = 9.37, p = .04, CFI = 0.95, TLI = 0.66, RMSEA = 0.14, χ

2Diff (2) = 8.81, p

< .05, indicating that the strength of the paths linking the two parental behaviors and

children’s EF at Time 2 were significantly different than those linking the two parental

behaviors and children’s VA at Time 2. In summary, our model showed that negative parent-

child interactions and parental scaffolding showed unique and specific associations with

individual differences in children’s EF but not with VA. In contrast the HLE and parental

MLU were related to both EF and VA. Parental EF was unrelated to either children’s EF or

VA in the final model.

Supplementary analysis. To confirm our findings we tested our first auto-regressive

longitudinal model using only those cases with complete data for all variables (N = 101).

Once again this model provided an acceptable fit to the data, χ2 (17) = 25.75, p = .08, CFI =

0.93, TLI = 0.92, RMSEA = 0.07. This model accounted for a similar proportion of the

variance in Time 2 EF (47%) and Time 2 VA (53%) as the full information model. Crucially,


23

each of the paths in this model showed the same pattern of findings as the full information

model. Importantly, negative parent-child interactions were uniquely and specifically related

to Time 2 EF, β = -.25, p < .01, 95%CI [-.43, -.07], but not Time 2 VA, β = -.10 p = .25,

95%CI [-.27, .07]. Furthermore, parental scaffolding showed unique and specific associations

with Time 2 EF, β = .20, p < .05, 95%CI [.03, .36], but not Time 2 VA, β = .11, p = .16,

95%CI [-.04, .27].

Discussion

This study included multiple detailed measures of parental behavior as well as

assessments of both parent EF and child EF across a 13-month period. Key questions

concerned whether positive and negative parental behaviors: (i) show unique associations

with child EF when considered in tandem; and (ii) show specific associations with child EF

(rather than being related to children’s cognitive ability in general). In relation to the first two

questions our findings revealed that negative parent-child interactions (characterized by

negative affect, criticism and control) showed a unique, specific inverse association with

child EF while parental scaffolding showed a unique, specific positive association with child

EF. In contrast, informal opportunities for learning (as measured by the Home Learning

Environment questionnaire) and parental language input were related to both EF and verbal

ability. Together, these findings support a differentiated model of parenting, in that each of

the measures of parental behavior was largely distinct and showed different patterns of

association with child EF and VA. Below, each set of findings is discussed in turn. In doing

so, we also outline the strengths and limitations of our study and offer some suggestions for

future research.

A Differentiated Model of Parental Influences on Children’s EF

The growth of evidence that EF plays a central role in children’s academic

achievement (e.g., Blair & Razza, 2007) and behavioral adjustment (e.g., Hughes, White,


24

Sharpen, & Dunn, 2000) has led to an explosion of research interest in family influences on

children’s EF. To date, with few exceptions (e.g., Bernier, et al., 2010; Hughes & Ensor,

2009), the great majority of studies on parental influences on children’s EF have focused on

single dimensions of parental behavior and have not examined the specificity of the effects of

both positive and negative parental behaviors on children’s EF.

Supporting a differentiated view of parental behavior (e.g., Carr & Pike, 2012;

Hughes & Ensor, 2009; Zheng et al., 2016) our results showed that parental behaviors can be

measured along distinct cognitive and affective dimensions. For instance the correlation

matrix shows that negative parent-child interaction was inversely related to parental talk

(MLU). That is, parents who engaged in extended discourse with their preschool child during

the free play task were not necessarily more adept at fine-tuning the level of their support

during the structured task or more likely to report providing their child with frequent learning

opportunities (e.g., reading, painting etc.). This in itself is a point worth emphasising, in that

it highlights the need for researchers to include a broad array of parenting measures in order

to capture a full picture of social influences on children’s cognitive development.

Our findings also have implications for educational and health professionals seeking

to improve developmental outcomes for at-risk children. For example, Merz, Landry,

Johnson, Williams and Jung (2016) have shown that an intervention designed to increase

responsiveness in foster parents caring for children aged 2.5 to 5 years had limited benefits

for children’s EF, in that improvements in the intervention group were restricted to the

youngest children in their sample. Coupled with findings from school-based interventions

that highlight the need for an incremental approach to ensure that children continue to receive

an appropriate level of challenge, our results indicate that family-based interventions to

promote EF in preschool children should adopt a multi-pronged approach that goes beyond


25

responsiveness to include a dual focus on promoting parents’ cognitive scaffolding and

reducing negative affect.

Parental Behaviors Show Both Overlapping and Specific Links With Children’s EF and

VA

Consistent with a domain-specific account of parental influences on children’s EF,

this study showed contrasting patterns of association between parental variables and

children’s ability. Specifically, parental scaffolding and negative parent-child interaction

showed specific and unique associations with child EF (in the expected directions) and

neither variable was associated with child VA. Confirming this, changes in model fit

indicated that these paths showed a significant contrast in strength across the two cognitive

domains (EF and VA). The specific link between negative parent-child interactions and

children’s EF is consistent with psychobiological accounts, which identify early negative

experiences as stressors and highlight the vulnerability of EF to stress (Blair & Raver, 2015).

Blair et al. (2011) have shown that the hypothalamic-pituitary-adrenal (HPA) axis

functioning, as measured by elevated salivary cortisol mediates the link between stressful

experiences and the emergence of EF. Future studies should investigate whether these

mechanisms are specific to EF.

At first glance, the specificity of the association between parental scaffolding and

child EF is surprising, as previous work has shown that language might play a mediating role

in the association between parental scaffolding and later child EF (Landry et al., 2002).

However, at least two between-study contrasts deserve note. The first of these is a clear

difference in what is meant by parental scaffolding. In Landry et al.’s (2002) study (and in

many other studies), the term ‘scaffolding’ encompasses a variety of supportive parental

behaviors, including praise and encouragement. By contrast, in the current study we used the

term ‘scaffolding’ much more tightly, to refer to the specific process of contingent shifting


26

(e.g., Carr & Pike, 2012). By this definition, scaffolding can be and often is non-verbal. For

example, parents might wordlessly remove all the non-edge pieces of a jigsaw to assist a

child who is struggling to complete the perimeter of the puzzle. Similarly, scaffolding also

includes reductions in the level of parental support in response to a child’s display of mastery.

The second between-study contrast concerns the measurement of EF. While Landry et

al. (2002) used children’s goal-directed behavior within the parent-child structured play

session to index child EF, the current study used four experimental tasks that did not require

verbal responses. Thus in this study both the independent and dependent variables were less

likely to entail verbal skills than the corresponding measures in the Landry et al. (2002)

study. Note also that our finding of a specific link between parental scaffolding and child EF

is consistent with the evidence from school-based interventions, which highlight the need for

incremental challenges that enable practice to be tailored to a child’s level of EF ability

(Diamond & Lee, 2011). Thus the specificity of the relations between parental scaffolding

and child EF also sheds light on the potential mechanisms that might underpin the

socialization of EF in early childhood.

While the home learning environment showed a general association with child EF and

VA, parental talk (as measured by the MLU) was associated in different directions with child

VA and child EF. The negative association between parental MLU and child EF is puzzling

but may reflect a child-driven effect: if children lack the EF skills to display autonomy,

parents may find it necessary to provide more detailed conversational support during free

play. Alternatively, a negative association between parental language input and children’s EF

could indicate that parents who talk too much might not be giving their children room to self-

regulate. Further work is needed to distinguish between these alternative accounts.

One strength of the current study was its inclusion of both direct observational coding

of parenting and self-report ratings of the home learning environment. This multi-measure


27

approach enabled us to assess whether parental responses to child behaviors predict variation

in gains in EF across a 13-month period, even when measures of the general home context

(parental education, home learning environment, MLU) are taken into account. Our findings

confirm the uniqueness of this association and also indicate the value of adopting a dual focus

on both positive and negative as well as cognitive and affective dimensions of parental

influence. The contrasting patterns of association between children’s EF and VA with

specific parental measures highlight the value of tailored interventions that focus on

particular aspects of parenting to promote distinct aspects of cognitive development.

Parental EF, Parent-Child Interactions and Children’s EF

In this study, negative parent-child interactions were related to low parental EF. This

is an important finding, as negative parent-child interactions are associated with a range of

long-term adverse outcomes including conduct problems (e.g.,Mills-Koonce, Willoughby,

Garrett-Peters, Wagner, & Vernon-Feagans, 2016), anxiety (e.g., Wood, McLeod, Sigman,

Hwang, & Chu, 2003) and depression (Cole et al., 2015). Understanding how EF skills can

constrain parental responses to difficult situations may prove useful in reducing parental

criticism or intrusiveness, particularly if parents attribute their responses to child

characteristics rather than to their own cognitive profiles.

At the same time, this result needs to be interpreted with caution as the association

between poor parental EF and negative parent-child interaction may simply reflect the

common influence of a third variable. One plausible candidate in this respect is depression,

which is associated with both reduced EF (e.g.,Austin, Mitchell, & Goodwin, 2001) and

negative parenting (Lovejoy, Graczyk, O'Hare, & Neuman, 2000). The lack of any measure

of parental wellbeing is a limitation of the current study and an obvious step for future

research is to examine the independence and interplay of parental depression and EF as

predictors of parental behaviors.


28

As noted in our introduction, genetic factors might also contribute to the association

between parental EF and negative parent-child interaction and so the lack of a genetically

sensitive design is a further limitation of the current study. That said, it is worth noting that

genetic variation is typically most salient for phenotypic individual differences that are

general and stable. In contrast, environmental factors typically show specific associations

with changes in a specific ability or trait (Kovas, Haworth, Dale, & Plomin, 2007). By

including a measure of child verbal ability in order to examine the specificity of associations

between parental measures and by assessing gains in child EF across two time-points, this

study minimized the extent to which the results might simply reflect genetic factors. Instead,

our findings support psycho-biological models of regulation (e.g., Blair & Raver, 2015) in

which stressful environments adversely affect the development of the prefrontal cortex, the

neural substrate for EF. Future research might therefore profit from the inclusion of measures

of children’s stress levels during observed parent-child interactions (e.g., Blair et al., 2011).

Highlighting the specificity of the association between parental EF and negative

parent-child interactions, our results showed that variation in parental EF was unrelated to

individual differences in contingent shifting. At first glance, this null result is surprising: after

all, contingent shifting is, by definition, a flexible response to changing situations and so

should provide an ecological example of EF in action. One explanation may be that there is a

‘transmission gap’ between competence and performance (Meins, Fernyhough, Johnson, &

Lidstone, 2006). That is, a parent with good EF might not utilize these skills when interacting

with his or her child, perhaps because they lack experience in supporting children’s goal-

directed activities. Exploring whether parental expertise moderates the association between

parental EF and contingent shifting would be an interesting avenue for future research. Note

that the parental EF measures adopted in the current study focused on inhibition and task

switching. Contingent shifting might therefore be related to other dimensions of EF not


29

measured in the present study (e.g., working memory, the ability to monitor and update one’s

actions in response to changing circumstances). Future work would therefore benefit from

incorporating a wider range of parental EF measures.

Finally our correlation table showed that, consistent with the general nature of genetic

effects (Asbury & Plomin, 2014), parental EF showed associations with both children’s EF

and VA that were similar in magnitude. Moreover, our models indicated that parental EF did

not account for gains in children’s EF above other factors such as the quality of parent-child

interactions. Conceptually these findings suggest that intergenerational associations in

cognitive ability are non-specific in nature and reflect a complex set of intervening variables

rather than a deterministic association. For example, an impulsive parent is likely to adopt

short-term strategies for managing children’s behavior, which in turn might constrain

children’s developing skills in goal-directed tasks.

Study Limitations

Although this study extends previous findings on the relations between parental

behavior and children’s EF, a number of limitations deserve note. Given the time-consuming

nature of observational coding our longitudinal study involved assessment of parental

behavior at just one time point. That is, parent-child interactions were studied at Time 1 only

and parental reports of the HLE and parental EF were assessed at Time 2 only. As a result,

our model did not include developmental associations from child characteristics to later

parenting behavior. While the impact of child behavior on parenting was included in our

rating of scaffolding, a recent cross-lagged longitudinal study suggests that there may be

bidirectional associations between EF and parental behavior (Blair, et al., 2014). While this

finding underscores the importance of including parental measures at more than one time

point (Zheng et al., 2016), it is worth stressing that our analyses controlled for stability in

children’s EF such that any observed relations indicate potential developmental influence on


30

EF (Menard, 2002) and despite their concurrent measurement, our findings highlight the

relative independence of parental and child EF scores. The timing and format of measurement

of parental behaviors might also account for the weak and non-significant correlations

between different aspects of parental behavior. That said, EF in adulthood shows high levels

of rank-order stability over time (e.g., Friedman et al., 2016) as do measures of the HLE

across early childhood (e.g., Lehrl, Kluczniok, & Rossbach, 2016). Moreover we identified

correlations between negative parent-child interactions (at Time 1) and parental EF (at Time

2). Future studies that incorporate parental measures at more than one time point will aid our

understanding of the relations between parental behavior and parental EF.

Conclusions

The results of this study indicate that including multiple direct observational measures

of parenting is labor well-spent, in that our findings confirm a multi-faceted model of

parenting (Zheng, et al., 2016). First, separate measures of parental behavior (i.e.,

positive/negative and cognitive/affective) show distinct patterns of association with

children’s EF. Second, our findings confirm the domain-specificity of parental influences.

That is, while scaffolding and negative parent-child interactions are both key predictors of

child EF, highlighting the diversity of parental influences on child EF, they were unrelated to

children’s VA. In contrast, the home learning environment was positively related to both

child outcomes and parental MLU was positively related to child VA but negatively related to

child EF. While unexpected, this last finding highlights the importance of going beyond

simple quantitative measures of talk to focus on the quality of parent-child interactions in

supporting children’s cognitive development (Hirsh-Pasek et al., 2015). By adopting multi-

dimensional measures of parental behavior in studies that include more diverse samples and

genetically sensitive designs, we hope that future research will expand our understanding of

positive and negative parental influences on children’s EF.


31

References

Asbury, K. & Plomin, R. (2014). G is for Genes: The Impact of Genetics on Education and

Achievement. Oxford, UK: Wiley Blackwell.

Austin, M., Mitchell, P., & Goodwin, G. (2001). Cognitive deficits in depression: Possible

implications for functional neuropathology. British Journal of Psychiatry, 178, 200-

206. DOI: 10.1192/bjp.178.3.200

Bernier, A., Carlson, S., & Whipple, N. (2010). From external regulation to self-regulation:

Early parenting precursors of young children's executive functioning. Child

Development, 81, 326-339. DOI: 10.1111/j.1467-8624.2009.01397.x

Blair, C., Granger, D., Willoughby, M., Mills-Koonce, R., Cox, M., Greenberg, M., Kivlighan,

K.T., Fortunato, C., & The FLP Investigators (2011). Salivary Cortisol Mediates Effects

of Poverty and Parenting on Executive Functions in Early Childhood. Child

Development, 82 1970-1984. DOI: 10.1111/j.1467-8624.2011.01643.x

Blair, C., & Raver, C. (2015). School readiness and self-regulation: a developmental

psychobiological approach. Annual Review of Psychology, 66, 711-731. DOI:

10.1146/annurev-psych-010814-015221

Blair, C., Raver, C., & Berry, D. (2014). Two approaches to estimating the effect of parenting

on the development of executive function in early childhood. Developmental

Psychology, 50, 554-565. DOI:10.1037/a0033647

Brown, T. (2015). Confirmatory Factor Analysis for Applied Research (Second ed.). London:

Guilford Press.

Carr, A., & Pike, A. (2012). Maternal scaffolding behavior: Links with parenting style and

maternal education. Developmental Psychology, 48, 543-551. DOI:10.1037/a0025888

Cole, D., Sinclair-McBride, K., Zelkowitz, R., Bilsk, S., Roeder, K., & Spinelli, T. (2015). Peer

Victimization and Harsh Parenting Predict Cognitive Diatheses for Depression in

https://dx.doi.org/10.1111%2Fj.1467-8624.2011.01643.x

https://dx.doi.org/10.1037%2Fa0033647

http://dx.doi.org/10.1037/a0025888


32

Children and Adolescents. Journal of Clinical Child and Adolescent Psychology, 13,

1537-4416 DOI: 10.1080/15374416.2015.1004679

Cragg, L., & Nation, K. (2007). Self-ordered pointing as a test of working memory in typically

developing children. Memory, 15, 526 - 535. DOI: 10.1080/09658210701390750

Cuevas, K., Deater-Deckard, K., Kim-Spoon, J., Wang, Z., Morasch, K., & Bell, M. (2014). A

longitudinal intergenerational analysis of executive functions during early childhood.

British Journal of Developmental Psychology, 32, 50-64. DOI:10.1111/bjdp.12021

Cuevas, K., Deater-Deckard, K., Kim-Spoon, J., Watson, A., Morasch, K., & Bell, M. (2014).

What's mom got to do with it? Contributions of maternal executive function and

caregiving to the development of executive function across early childhood.

Developmental Science, 17, 224-238. DOI: 10.1111/desc.12073

Davidov, M. (2013). The socialization of self-regulation from a domains perspective. In G.

Seebaß, M. Schmitz. & P.M. Gollwitzer (Eds). Acting Intentionally and its Limits:

Individuals, Groups, Institutions (pp. 223-244). Berlin, Germany: De Gruter.

Davidson, M., Amso, D., Anderson, L., & Diamond, A. (2006). Development of cognitive

control and executive functions from 4 to 13 years: Evidence from manipulations of

memory, inhibition, and task switching. Neuropsychologia,, 44, 2037-2078. DOI:

10.1016/j.neuropsychologia.2006.02.006

Deater-Deckard, K., Pylas, M., & Petrill, S. (1997). Parent-Child Interactive System

(PARCHISY). London: Institute of Psychiatry, King's College London.

Devine, R.T., Bignardi, G., & Hughes, C. (2016). Executive Function Mediates the Relations

between Parental Behaviors and Children’s Early Academic Ability. Frontiers in

Psychology. 7,1902. doi: 10.3389/fpsyg.2016.01902

http://doi.org/10.1080/09658210701390750

https://dx.doi.org/10.1111%2Fbjdp.12021

https://dx.doi.org/10.1111%2Fdesc.12073

https://dx.doi.org/10.1016%2Fj.neuropsychologia.2006.02.006


33

Devine, R.T. & Hughes, C. (2014). Relations between false belief understanding and executive

function in early childhood: a meta-analysis. Child Development, 85, 1777-1794. DOI:

10.1111/cdev.12237

Diamond, A., & Lee, K. (2011). Interventions shown to aid executive function development in

children 4 to 12 years old. Science, 333, 959-964. DOI: 10.1126/science.1204529

Eley, T.C. & Stevenson, J. (1999). Exploring covariation between anxiety and depression

symptoms: A genetic analysis of the effects of age and sex. Journal of Child

Psychology and Psychiatry, 40, 1273 - 1282. DOI: 10.1111/1469-7610.00543

Fay-Stammbach, T., Hawes, D.J. & Meredith, P. (2014). Parenting influences on executive

function in early childhood: A review. Child Development Perspectives, 8, 258-264.

DOI: 10.1111/cdep.12095

Friedman, N., Miyake, A., Altamirano, L., Corley, R., Young, S., Rhea, S., & Hewitt, J. (2016).

Stability and change in executive function abilities from late adolescence to early

adulthood: A longitudinal twin study. Developmental Psychology, 52, 326-340. DOI:

10.1037/dev0000075.

Garon, N., Bryson, S., & Smith, I. (2008). Executive function in preschoolers: A review using

an integrative framework. Psychological Bulletin, 134, 31-60. DOI: 10.1037/0033-

2909.134.1.31.

Gerstadt, C., Hong, Y., & Diamond, A. (1994). The relationship between cognition and action:

performance of children 31/2 - 7 years old on a stroop-like day-night test. Cognition,

53, 129-153. DOI: 10.1016/0010-0277(94)90068-X

Hammond, S., Muller, U., Carpendale, J., Bibok, M., & Lieberman-Finestone, D. (2012). The

effects of parental scaffolding on preschoolers’ executive function. Developmental

Psychology, 48, 271-281. DOI: 10.1037/a0025519

https://dx.doi.org/10.1126%2Fscience.1204529

https://doi.org/10.1016/0010-0277(94)90068-X


34

Hertzog, C., & Nesselroade, J. (2003). Assessing psychological change in adulthood.

Psychology and Aging, 18, 639-657. DOI: 10.1037/0882-7974.18.4.639

Hirsh-Pasek, K., Adamson, L., Bakeman, R., Owen, M., Golinkoff, R., Pace, A., Yust, P.K., &

Suma, K. (2015). The Contribution of Early Communication Quality to Low-Income

Children’s Language Success. Psychological Science, 26, 1071-1083. DOI:

10.1177/0956797615581493

Hoff, E. (2006). How social contexts support and shape language development. Developmental

Review,, 26 55-88. DOI: 10.1016/j.dr.2005.11.002

Hughes, C., & Ensor, R. (2005). Theory of Mind and Executive Function in 2-year-olds: A

family affair? Developmental Neuropsychology, 28, 645-668. DOI:

10.1207/s15326942dn2802_5

Hughes, C., & Ensor, R. (2007). Positive and Protective: Effects of Early Theory of Mind on

Preschool Problem Behaviors. Journal of Child Psychology and Psychiatry, 48, 1025-

1032. DOI: 10.1111/j.1469-7610.2007.01806.x

Hughes, C., & Ensor, R. (2009). How do families help or hinder the development of executive

function? New Directions in Child and Adolescent Psychiatry: Special Issue on social

interaction and the development of executive function, 123, 35-50. DOI:10.1002/cd.234.

Hughes, C., Graham, A., & Grayson, A. (2005). Executive Function in Childhood:

Development and disorder. In J. Oates (Ed.), Cognitive Development (pp. 205-230):

Open University Press.

Hughes, C., Roman, G., Hart, M., & Ensor, R. (2013). Does Maternal Depression Predict

Young Children’s Executive Function? A 4-year Longitudinal Study. Journal of Child

Psychology and Psychiatry, 54, 169-177. DOI: 10.1111/jcpp.12014

Hughes, C., White, A., Sharpen, J., & Dunn, J. (2000). Antisocial, angry and unsympathetic:

‘Hard to manage’ preschoolers’ peer problems, and possible social and cognitive

https://doi.org/10.1037/0882-7974.18.4.639

http://dx.doi.org/10.1207/s15326942dn2802_5


35

influences. Journal of Child Psychology and Psychiatry, 41, 169-179. DOI:

10.1111/1469-7610.00558

Kolb, B., Mychasiuk, R., Muhammad, A., Li, Y., Frost, D., & Gibb, R. (2012). Experience and

the developing prefrontal cortex. Proceedings of the National Academy of Sciences of

the United States of America, 109 17186-17193. DOI: 10.1073/pnas.1121251109

Kovas, Y., Haworth, C., Dale, P., & Plomin, R. (2007). The Genetic and Environmental

Origins of Learning Abilities and Disabilities in the Early School Years (Vol. 72).

Boston: Blackwell.

Lagattuta, K., Sayfan, L., & Monsour, M. (2011). A new measure for assessing executive

function across a wide age range: Children and adults find happy-sad more difficult

than day-night. Developmental Science, 14, 481-489. DOI:10.1111/j.1467-

7687.2010.00994.x

Landry, S.H., Miller-Loncar, C.L., Smith, K.E., Swank, P.R. (2002). The role of early

parenting in children's development of executive processes. Developmental

Neuropsychology, 21, 15-41. DOI: 10.1207/S15326942DN2101_2

Lehrl, S., Kluczniok, K., & Rossbach, H. (2016). Long-term associations of preschool

education: the predictive role of preschool quality for the development of mathematical

skills through elementary school. Early Child Research Quarterly, 36, 475-488. DOI:

10.1016/j.ecresq.2016.01.013

Leve, L., DeGarmo, D., Bridgett, D., Neiderhiser, J., Shaw, D., Harold, G., Natsuaki, M.N., &

Reiss, D. (2013). Using an adoption design to separate genetic, prenatal, and

temperament influences on toddler executive function. Developmental Psychology, 49,

1045-1057. DOI:10.1037/a0029390

https://dx.doi.org/10.1073%2Fpnas.1121251109

http://dx.doi.org/10.1207/S15326942DN2101_2

https://dx.doi.org/10.1037%2Fa0029390


36

Lovejoy, M., Graczyk, P., O'Hare, E., & Neuman, G. (2000). Maternal depression and

parenting behavior: A meta-analytic review. Clinical Psychology Review, 20, 561-592.

DOI: 10.1016/S0272-7358(98)00100-7

Mackenzie, M., Nicklas, E., Waldfogel, J., & Brooks-Gunn, J. (2012). Corporal Punishment

and Child Behavioral and Cognitive Outcomes through 5Years of Age: Evidence from a

Contemporary Urban Birth Cohort Study. Infant and Child Development,, 21, 3-33.

DOI: 10.1002/icd.758

Meins, E. (1997). Security of attachment and maternal tutoring strategies: Interaction within the

zone of proximal development. British Journal of Developmental Psychology, 15, 129-

144. DOI: 10.1111/j.2044-835X.1997.tb00730.x

Meins, E., Fernyhough, C., Johnson, F., & Lidstone, J. (2006). Mind-mindedness in children:

Individual differences in internal-state talk in middle a childhood. British Journal of

Developmental Psychology, 24, 181-196. DOI: 10.1348/026151005X80174

Melhuish, E., Sylva, K., Sammons, P., Siraj-Blatchford, I., Taggart, B., Phan, M., & Malin, A.

(2008). Preschool influences on mathematics achievement. Science, 321, 1161-1162.

DOI:10.1126/science.1158808

Menard, S. (2002). Longitudinal Research (Second ed.). London: Sage.

Merz, E., Harlé, K., Noble, K., & Mccall, R. (2016). Executive Function in Previously

Institutionalized Children. Child Development Perspectives, 10, 105-110. DOI:

10.1111/cdep.12170

Merz, E., Landry, S., Johnson, U., Williams, J., & Jung, K. (2016). Effects of a responsiveness-

focused intervention in family child care homes on children's executive function. Early

Childhood Research Quarterly, 34, 128-139. DOI: 10.1016/j.ecresq.2015.10.003

Mills-Koonce, W., Willoughby, M., Garrett-Peters, P., Wagner, N., & Vernon-Feagans, L.

(2016). The interplay among socioeconomic status, household chaos, and parenting in

https://doi.org/10.1016/S0272-7358(98)00100-7

https://dx.doi.org/10.1016%2Fj.ecresq.2015.10.003


37

the prediction of child conduct problems and callous-unemotional behaviors.

Development and Psychopathology, 28, 757-771. DOI: 10.1017/S095457941600029

Miyake, A., Friedman, N., Emerson, M., Witzki, A., Howerter, A., & Wager, T. (2000). The

unity and diversity of executive functions and their contributions to complex "frontal

lobe" tasks: A latent variable analysis. Cognitive Psychology, 41, 49-100. DOI:

10.1006/cogp.1999.0734

Mulvaney, M., McCartney, K., Bub, K., & Marshall, N. (2006). Determinants of dyadic

scaffolding and cognitive outcomes in first graders. Parenting: Science and Practice, 6,

297–320. DOI:10.1207/s15327922par0604_2

Noble, K., Norman, M. F., & Farah, M. (2005). Neurocognitive correlates of socioeconomic

status in kindergarten children. Developmental Science, 8, 74-87. DOI: 10.1111/j.1467-

7687.2005.00394.x

Pratt, M., Green, D., MacVicar, J., & Bountrogianni, M. (1992). The mathematical parent:

Parental scaffolding, parenting style, and learning outcomes in long-division

mathematics homework. Journal of Applied Developmental Psychology, 13, 17-34.

DOI: 10.1016/0193-3973(92)90003-Z

Rust, J. (2003). The Wechsler Preschool and Primary Scale of Intelligence (WPPSI III UK)

(Third ed.). London: Harcourt Assessments.

Smetana, J.G. (2017). Current research on parenting styles, dimensions and beliefs. Current

Opinon in Psychology, 15, 19 - 25. DOI: 10.1016/j.copsyc.2017.02.012

Son, S., & Morrison, F. (2010). The nature and impact of changes in home learning

environment on development of language and academic skills in preschool children.

Developmental Psychology, 46, 1103-1118. DOI: 10.1037/a0020065

http://dx.doi.org/10.1207/s15327922par0604_2

https://doi.org/10.1111/j.1467-7687.2005.00394.x

https://doi.org/10.1111/j.1467-7687.2005.00394.x

https://doi.org/10.1016/0193-3973(92)90003-Z


38

Tausczik, Y., & Pennebaker, J. (2010). The psychological meaning of words: LIWC and

computerized text analysis methods. Journal of Language and Social Psychology, 29,

24-54. DOI: 10.1177/0261927X09351676

Ursache, A., & Noble, K. (2016). Neurocognitive development in socioeconomic context:

Multiple mechanisms and implications for measuring socioeconomic status.

Psychophysiology, 53, 71-82. DOI: 10.1111/psyp.12547

Vygotsky, L. (1978). Mind in society: The development of higher psychological processes.

Cambridge MA: Harvard University Press.

Wiebe, S.A., Espy, K.A. & Charak, D. (2008). Using confirmatory factor analysis to

understand executive control in preschool children: I. Latent structure. Developmental

Psychology, 44, 575 - 87. DOI: 10.1037/0012-1649.44.2.575

Wood, D., Bruner, J., & Ross, G. (1976). The role of tutoring in problem solving. Journal of

Child Psychology and Psychiatry, 17, 89-100. DOI:10.1111/j.1469-

7610.1976.tb00381.x

Wood, D., & Middleton, D. (1975). A study of assisted problem-solving. British Journal of

Psychology, 66, 181-191. DOI: 10.1111/j.2044-8295.1975.tb01454.x

Wood, J., McLeod, B., Sigman, M., Hwang, W., & Chu, B. (2003). Parenting and childhood

anxiety: Theory, empirical findings, and future directions. Journal of Child Psychology

and Psychiatry, 44, 134-151. DOI: 10.1111/1469-7610.00106

Zelazo, P. (2006). The Dimensional Change Card Sort (DCCS): a method of assessing

executive function in children. Nature Protocols, 1, 297 - 301. DOI:

10.1038/nprot.2006.46

Zheng, Y., Pasalich, D., Oberth, C., McMahon, R., & Pinderhughes, E. (2016). Capturing

Parenting as a Multidimensional and Dynamic Construct with a Person-Oriented

Approach. Prevention Science, 1-11. DOI: 10.1007/s11121-016-0665-0

https://doi.org/10.1037/0012-1649.44.2.575


39

Figure 1. Standardized Robust Maximum Likelihood Estimates for Longitudinal Model.

Note. **p < .01. *p < .05. Only significant paths are depicted. Full model results are shown in Table S1. EF = Executive Function. VA = Verbal

Ability. IQ = General Cognitive Ability. Scaffold = Parental Scaffolding. Negative = Negative Parent-Child Interactions. HLE = Home Learning

Environment. MLU = Parent Mean Length of Utterance. PEF = Parent Executive Function. PED = Parent Education. T1 = Time 1. T2 = Time 2.


40

Table 1. Descriptive Statistics

Time 1 Time 2

M (SD) 95% CI Range M (SD) 95% CI Range

Child Measures

Happy-Sad Stroop 13.35 (6.30) 12.17, 14.53 0 – 20 17.24 (2.42) 16.77, 17.71 7 – 20

DCCS Post-Switch 3.74 (2.50) 3.29, 4.19 0 – 6 5.36 (1.63) 5.04, 5.67 0 – 6

Self-Ordered Pointing Task 0.13 (0.06) 0.12, 0.14 0 - 0.28 0.08 (0.04) 0.07, 0.09 0 – 0.22

DCCS Border Game - - - 6.57 (2.86) 6.02, 7.12 0 – 12

Day-Night Stroop - - - 17.26 (1.98) 16.87, 17.65 12 – 20

Receptive Vocabulary 23.04 (5.12) 22.12, 23.96 8 – 31 27.08 (3.82) 26.34, 27.82 17 – 34

Object Assembly 16.71 (8.24) 15.22, 18.20 2 – 35 - - -

Parent Measures

Parental Control 49.94 (7.89) 48.44, 51.44 41.26 – 82.49 - - -

Parental Scaffolding 0.42 (0.20) 0.38, 0.46 0 – 1 - - -

Mean Length of Utterance 5.25 (0.99) 5.05, 5.45 3.37 – 9.10 - - -

Home Learning Environment - - - 30.44 (9.78) 28.56, 32.32 8 – 49

Arrows Efficiency (trial/s) - - - 1.14 (0.57) 1.03, 1.25 0 – 2.15

Hearts/Flowers Efficiency (trial/s) - - - 0.99 (0.54) 0.88, 1.09 0 – 2.39


41

Table 2. Sample Correlations.

Variable 1 2 3 4 5 6 7 8 9 10 11 12

1. Child Executive Function T1 -

2. Child Executive Function T2 .49** -

3. Child Verbal Ability T1 .37** .42** -

4. Child Verbal Ability T2 .47** .48** .59** -

5. Child Cognitive Ability T1 .43** .42** .39** .37** -

6. Child Age (Concurrent) .55** .51** .51** .46** .52** -

7. Child Gender -.09 -.15 -.16 -.09 .09 .05 -

8. Negative Parent-Child

Interaction

-.09 -.29** -.21* -.25* -.25** -.07 .19* -

9. Parental Scaffolding .18+ .26** .10 .20* .02 .02 -.06 -.12 -

10. HLE .03 .10 .04 .18+ -.05 -.01 -.22* .18* .05 -

11. MLU .10 .001 .06 .32** .20* .07 .02 -.30** .01 .06 -

12. Parent Executive Function .18+ .22* .24* .16 .23* .02 -.06 -.22* -.06 -.06 -.03 -

13. Parent Education .16 .15 .15 .12 .12 -.04 .06 .05 .20* .08 .12 .07

Note. **p < .01. *p < .05. +p < .10. T1 = Time 1. T2 = Time 2. 103 < N < 117. T1 = Time 1. T2 = Time 2. HLE = Home Learning Environment.

MLU = Mean Length of Utterance.

university of birmingham for better or for worse?

Documents